From: "zhaoyang.huang" <zhaoyang.huang@unisoc.com>
To: Andrew Morton <akpm@linux-foundation.org>,
Matthew Wilcox <willy@infradead.org>,
Jens Axboe <axboe@kernel.dk>, Tejun Heo <tj@kernel.org>,
Josef Bacik <josef@toxicpanda.com>,
Baolin Wang <baolin.wang@linux.alibaba.com>, <linux-mm@kvack.org>,
<linux-block@vger.kernel.org>, <linux-kernel@vger.kernel.org>,
<cgroups@vger.kernel.org>,
Zhaoyang Huang <huangzhaoyang@gmail.com>, <steve.kang@unisoc.com>
Subject: [RFC PATCH 0/2] introduce budgt control in readahead
Date: Thu, 9 May 2024 10:39:35 +0800 [thread overview]
Message-ID: <20240509023937.1090421-1-zhaoyang.huang@unisoc.com> (raw)
From: Zhaoyang Huang <zhaoyang.huang@unisoc.com>
Over-limit bw value is observed during fio test in the throttling group
which caused by over-sized bio as there is no control on ra->size during
readahead. This series patches would like to introduce the helper
function to provide the bytes limit and apply it on readahead.
Please find below for the fio test result on v6.6 which presents 2%-10%
improvement for BW and lat. Besides, we can also observed stable BW
instantaneous value during the test.
blkio.throttle.read_bps_device = 1MB/s
before: read: IOPS=223, BW=894KiB/s (915kB/s)(175MiB/200919msec)
after : read: IOPS=239, BW=960KiB/s (983kB/s)(153MiB/163105msec)
before: clat (usec): min=4, max=16795k, avg=4468.74, stdev=265746.14
lat (usec): min=6, max=16795k, avg=4470.57, stdev=265746.14
after : clat (usec): min=11, max=209193, avg=4105.22, stdev=27188.04
lat (usec): min=16, max=209197, avg=4120.03, stdev=27188.04
blkio.throttle.read_bps_device = 10MB/s
before: read: IOPS=2380, BW=9524KiB/s (9752kB/s)(1007MiB/108311msec)
after : read: IOPS=2438, BW=9754KiB/s (9989kB/s)(1680MiB/176405msec)
before: clat (usec): min=4, max=201817, avg=399.58, stdev=8268.85
lat (usec): min=6, max=201819, avg=402.10, stdev=8268.85
after : clat (usec): min=4, max=2494.6k, avg=412.72, stdev=25783.51
lat (usec): min=6, max=2494.6k, avg=414.48, stdev=25783.51
Zhaoyang Huang (2):
block: introduce helper function to calculate bps budgt
mm: introduce budgt control in readahead
block/blk-throttle.c | 44 ++++++++++++++++++++++++++++++++++++++
include/linux/blk-cgroup.h | 10 +++++++++
mm/readahead.c | 33 ++++++++++++++++++++--------
3 files changed, 78 insertions(+), 9 deletions(-)
--
2.25.1
next reply other threads:[~2024-05-09 2:41 UTC|newest]
Thread overview: 12+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-05-09 2:39 zhaoyang.huang [this message]
2024-05-09 2:39 ` [RFC PATCH 1/2] block: introduce helper function to calculate bps budgt zhaoyang.huang
2024-05-09 2:39 ` [RFC PATCH 2/2] mm: introduce budgt control in readahead zhaoyang.huang
2024-05-09 3:15 ` Matthew Wilcox
2024-05-10 2:43 ` Zhaoyang Huang
2024-05-10 3:18 ` Matthew Wilcox
2024-05-11 7:35 ` Zhaoyang Huang
2024-05-14 2:37 ` Zhaoyang Huang
2024-05-09 12:39 ` Christoph Hellwig
2024-05-10 3:06 ` Zhaoyang Huang
2024-05-10 4:14 ` Matthew Wilcox
2024-05-10 7:08 ` Zhaoyang Huang
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20240509023937.1090421-1-zhaoyang.huang@unisoc.com \
--to=zhaoyang.huang@unisoc.com \
--cc=akpm@linux-foundation.org \
--cc=axboe@kernel.dk \
--cc=baolin.wang@linux.alibaba.com \
--cc=cgroups@vger.kernel.org \
--cc=huangzhaoyang@gmail.com \
--cc=josef@toxicpanda.com \
--cc=linux-block@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=steve.kang@unisoc.com \
--cc=tj@kernel.org \
--cc=willy@infradead.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).